Title :
Interval arithmetic backpropagation
Author :
Hernández, C.A. ; Espf, J. ; Nakayama, K. ; Fernández, M.
Author_Institution :
Dept. of Comput. & Electron., Valencia Univ., Spain
Abstract :
Presents an extension of the backpropagation learning algorithm by using interval arithmetic. The proposed algorithm represents a generalization of backpropagation and contains backpropagation as a particular case. This new algorithm permits the use of training samples and targets which can be indistinct points and intervals. Among the possible applications of this algorithm, the authors report its usefulness to integrate expert\´s knowledge and experimental samples and also its ability to handle "don\´t care attributes" in a simple and natural way in comparison with backpropagation. It also adds flexibility to the codification of inputs and outputs.
Keywords :
backpropagation; neural nets; backpropagation learning algorithm; don´t care attributes; interval arithmetic backpropagation; Arithmetic; Backpropagation algorithms; Cost function; Equations; Mean square error methods; Multi-layer neural network; Neural networks; Neurons; Transfer functions;
Conference_Titel :
Neural Networks, 1993. IJCNN '93-Nagoya. Proceedings of 1993 International Joint Conference on
Print_ISBN :
0-7803-1421-2
DOI :
10.1109/IJCNN.1993.713935